Have questions? Leave your message here or Schedule a quick call with our manager now

Model Context Protocol · New

Connect AI agents
to your store

Claude, Cursor, VS Code, and other MCP clients read and write orders, products, and customers through the API you already use. No integration code required.

Overview

API2Cart MCP Server lets AI agents work with your connected e-commerce platforms through the same unified API you already use.

The agent authenticates with a scoped access token, discovers the tools it’s allowed to use, and reads or writes orders, products, and customers within the permissions you set.

No integration project. Just a conversation.

01

Works everywhere

Claude, Cursor, VS Code, OpenAI Codex, Google Antigravity, LM Studio.

02

70+ platforms

Orders, products, and customers — one API across every store you connect.

03

Nothing to maintain

No custom connectors, no per-platform SDKs.

Config

Paste one block. You’re connected.

Generate a token, drop the config into your AI client, and start asking. No custom integration code.

  • Ready-made config for Claude, Cursor, VS Code, and more
  • Standard Bearer auth — treat tokens like passwords
  • Works over Streamable HTTP, no local server to run
MCP client config, pasted and connected
Token scope controls
Security

You decide exactly what the AI can touch

Every token is scoped — not a blanket API key handed to an agent.

  • Restrict a token to specific connections, integrations, or API methods
  • Hide raw connection keys so the agent only ever references connections by ID
  • Set a token lifetime from 1 hour to 180 days, and revoke it instantly
No-Code

No developer required

Store owners can connect an AI assistant themselves and start asking it questions in plain language — no ticket, no code.

  • “Show me today’s unfulfilled orders”
  • “Update the price on this product”
  • “Find every customer who ordered X last month”
Store owner asking their AI agent a question
mcp-config.json

The full configuration

{
  "mcpServers": {
    "api2cart": {
      "url": "https://mcp.api2cart.com",
      "headers": { "Authorization": "Bearer ****" }
    }
  }
}
Runs on
Claude · Cursor · VS Code · OpenAI Codex · Google Antigravity · LM Studio

Frequently asked questions

The things people ask before connecting an agent to their store.

Is my store data safe if I connect an AI agent?

Yes. Every token is scoped to exactly what you allow — specific connections, specific integrations, and specific API methods. An agent connected with a restricted token cannot reach anything outside that scope.

Which AI assistants are supported?

Any MCP-compatible client, including Claude Code, Cursor, VS Code, OpenAI Codex, Google Antigravity, and LM Studio.

Do I need to know how to code?

No. Generating a token and connecting a client takes a few clicks in your dashboard.

Can I limit what the AI can see or do?

Yes. Restrict a token to specific store connections, integrations, and API methods, and optionally hide raw connection keys.

How long does a token stay active?

You choose: 1 hour, 8 hours, 24 hours, 7 days, 30 days, or a custom duration up to 180 days. Revoke any token instantly.

Does this work with small or local models?

Yes — tested successfully with small local LLMs such as Gemma 4 12B (Q4).

Ready to connect your first AI agent?

Full documentation: How to Use API2Cart MCP. Announcement and use-case walkthroughs: for developers and for store owners.

Is my store data safe if I connect an AI agent?

Yes. Every token is scoped to exactly what you allow — specific connections, specific integrations, and specific API methods. An agent connected with a restricted token cannot reach anything outside that scope.

Which AI assistants are supported?

Any MCP-compatible client, including Claude Code, Cursor, VS Code, OpenAI Codex, Google Antigravity, and LM Studio.

Do I need to know how to code?

No. Generating a token and connecting a client takes a few clicks in your dashboard.

Can I limit what the AI can see or do?

Yes. Restrict a token to specific store connections, integrations, and API methods, and optionally hide raw connection keys.

How long does a token stay active?

You choose: 1 hour, 8 hours, 24 hours, 7 days, 30 days, or a custom duration up to 180 days. Revoke any token instantly.

Does this work with small or local models?

Yes — tested successfully with small local LLMs such as Gemma 4 12B (Q4).